Pellets/block bentonite barriers: Laboratory study of their evolution upon hydration

Some deep geological nuclear waste disposal concepts are considering the installation around the waste canisters of bentonite barriers made up of two components: highly compacted blocks to support the canister and a granular buffer material consisting of bentonite pellets to fill the rest of the disposal gallery. The initial characteristics of these two components in terms of dry density, water content and microstructure are very different, but some, yet scarce, experimental evidence showed that saturation with groundwater will tend to homogenise the barrier. Nevertheless, the combined use of pellets and blocks in the same section of the barrier introduces difficulties in the understanding and modelling of the system performance that have to be addressed. With the aim of obtaining qualitative and quantitative information about the hydro-mechanical evolution of initially inhomogeneous bentonite samples (pellets/blocks) upon hydration under isothermal conditions, a series of laboratory tests in decimetre-scale cells was carried out. In most tests hydration took place through the pellets part, but the effect of hydrating through the block was also checked. As well, tests were performed either under low water injection pressure or under low water inflow rate. The results highlighted the role of the large macropores on the initial stages of hydration when this took place through the pellets, and also of the influence of water availability on its redistribution and on pressure build up. Pressure developed for lower degrees of saturation and water spread more homogeneously through the two sample components if hydration was slow. After full saturation differences in dry density across the two components remained, but they were softer if saturation took place at a slow pace, either because of the low water inflow rate or because it occurred through the less permeable block. The microstructure of the bentonite in the two components was very different even after full saturation: although there was an overall trend to pore size homogenisation towards smaller sizes over time, the ratio of microstructural void ratio to macrostructural void ratio kept lower in the pellets than in the block. As well, the number of water layers in the interlayer was higher in pellets samples than in block samples. The research also highlighted the importance of the right assessment of macroporosity (particularly pores larger than 550 μm) from mercury intrusion porosimetry tests and the lik